Influence of Membrane Cathode Assembly Composition and Fabrication Methods on Proton Exchange Membrane Electrolyzers for Electrochemical Toluene Hydrogenation

Article information

Abstract

INTRODUCTION

The global increase in energy consumption, resulting in heightened carbon emissions, is accelerating global warming, which remains a critical issue that humanity must address [1,2]. Despite efforts to shift towards cleaner energy sources, fossil fuels still account for 80% of global energy production, leading to significant carbon emissions [3]. In response, many countries are exploring various energy production methods to move away from the carbon-dependent paradigm. Among these, renewable energy is the most environmentally friendly and produces no carbon emissions [4]. However, the intermittent nature of renewable energy requires efficient storage solutions, where green hydrogen produced via water electrolysis has been identified as a potential solution by many experts [5,6]. Despite its promise, hydrogen is challenging to store owing to the small and light nature of hydrogen molecules, which require considerable energy for storage. Consequently, extensive research has been conducted to develop efficient storage methods [7]. Liquid organic hydrogen carriers (LOHCs) have emerged as a promising solution owing to their ease of storage and transport as well as their high stability at room temperature [8,9]. Methylcyclohexane (MCH), a hydrogenated form of toluene, is particularly noteworthy among the LOHCs. Despite its high dehydrogenation temperature and lower volumetric hydrogen storage capacity compared with other carriers such as liquid hydrogen or ammonia, MCH is manageable and stable across a wide temperature range (–95 to 111°C), has a low toxicity, and uses the existing oil infrastructure [10,11].

Traditionally, the hydrogenation of toluene to MCH involves a two-step process. First, green hydrogen is produced via water electrolysis, which is then stored in toluene via an exothermic reaction that results in thermal loss (205 kJ mol^–1) [12,13]. The toluene hydrogenation also requires relatively high pressures and temperatures. To address these challenges, recent studies have focused on electrochemically hydrogenating toluene, which can be performed in a single step in a zero-gap electrolyzer using a solid polymer electrolyte [14,15]. In this system, toluene is electrochemically reduced at the cathode by reacting with the protons produced from the electrochemical oxidation of water at the anode. The electrochemical reactions occurring in a zero-gap electrolyzer with a proton exchange membrane (PEM) are as follows [16]:

where E^o is the standard reduction potential. As these reactions show, electrochemical toluene hydrogenation requires a cell voltage of 1.08 V or more, theoretically lowering the barrier to reaction initiation compared with water electrolysis, which requires a cell voltage of 1.23 V or more [17]. Furthermore, these reactions usually occur at relative low pressures and temperatures (e.g., 1 bar and 60°C), saving input energy.

Despite these advantages, research on the electrochemical hydrogenation of toluene remains limited. The basic configuration of a zero-gap electrolyzer (PEM electrolyzer) used in toluene hydrogenation is similar to that of a PEM water electrolyzer. However, owing to the organic nature of toluene and its role as a reactant at the cathode, efficient toluene hydrogenation requires a different approach from aqueous-based electrolyzers, particularly concerning the cathode. During electrochemical toluene hydrogenation, protons move from the anode to the cathode through the PEM, leading to water crossover, which disrupts toluene mass transfer at the cathode. This crossover can also trigger the hydrogen evolution reaction (HER) at the cathode, thereby reducing the current efficiency of toluene hydrogenation. Although several researchers have conducted basic studies on these phenomena [18–20], investigations into cathode fabrication methods and components that directly influence these processes in PEM electrolyzers are lacking. As a preliminary step towards understanding this complexity, we investigated how the cathode catalyst loading, ionomer content in the catalyst, the catalyst layer formation method, and introduction of a microporous layer (MPL) impact the electrochemical hydrogenation of toluene in the unit cell of a PEM electrolyzer. Using electrochemical impedance spectroscopy (EIS), we evaluated the performance of unit cells with various compositions of membrane cathode assemblies (MCAs) for toluene hydrogenation and discussed the factors influencing the results.

EXPERIMENTAL

Electrochemical measurements

Electrochemical measurements were conducted in a unit cell using an electrochemical analyzer (MULTI AUOTOLAB M204, Metrohm) equipped with a current amplifier (BOOSTER10A, Metrohm). The unit cell configuration is illustrated in Fig. 1. During the experiments, a 1 M aqueous solution of H₂SO₄ was supplied to the anode, and liquid toluene was supplied to the cathode, both at a flow rate of 5 mL min^–1. The cell was operated at 60°C. Cell polarization curves were obtained by scanning the cell voltage from 0.9 to 2.2 V at a scan rate of 0.01 V s^–1. Ohmic resistance (R_Ω) and charge transfer resistance (R_ct) were measured using EIS at an applied voltage of 1.6 V with a perturbation of 10 mV, within a frequency range of 0.1 to 100 Hz. To obtain iR-corrected polarization curves, the cell voltage was measured at constant current densities ranging from 0.02 to 0.50 A cm^–2, in increments of 0.01 A cm^–2. Voltage readings were averaged over a 60-second interval for each current density. EIS measurements were also performed under each current density condition to determine the high-frequency resistance (HFR) corresponding to R_Ω. The HFR was obtained from the high-frequency x-intercept of the Nyquist plot. The iR-corrected polarization curves were subsequently derived using the HFR values.

Fig. 1.

Schematic of the unit cell used for electrochemical toluene hydrogenation.

The current efficiency for toluene hydrogenation was determined during the operation of the unit cell at a constant current density of 0.10 A cm^–2. The catalyst loading on the cathode of the unit cell was 0.05 mg_PtRu cm^–2. A sample of the toluene and product MCH mixture was taken from the outlet of the unit cell, and the concentration of MCH in the mixture was measured using gas chromatography/mass spectrometry (ISQ LT, ThermoScientific). The current efficiency (CE) was then calculated using the following equation:

(3) CE=100×(nFQ/iA)×CMCH

where n is the number of electrons transferred (6), F is the Faraday constant (96485 C mol^–1), Q is the volumetric flow rate, i is the applied current density, A is the geometric electrode area, and C_MCH is the measured concentration of MCH.

RESULTS AND DISCUSSION

Fig. 2a–c presents the cell polarization curves of MCAs fabricated via the CCS method, with varying catalyst loadings and ionomer contents. The PTL used in theses MCAs was carbon paper without an MPL. The polarization curves exhibit significant performance variation, depending on the catalyst loading and ionomer content. For catalyst loadings of 0.5, 1.0, and 1.5 mg_PtRu cm^–2, the current density increased with rising ionomer content and subsequently decreased. For instance, at a catalyst loading of 1.0 mg_PtRu cm^–2 (Fig. 2c), the current density across the entire cell voltage range increased with ionomer content up to 20 wt% and then began to decrease at 30 wt%. Notably, the optimal ionomer content for maximum performance varied with catalyst loading. At a catalyst loading of 0.5 mg_PtRu cm^–2 (Fig. 2a), the highest current density of 0.141 A cm–² was achieved at 1.95 V with an ionomer content of 10 wt%. Meanwhile, higher catalyst loadings of 1.0 and 1.5 mg_PtRu cm^–2 yielded current densities of 0.298 and 0.421 A cm–² at 20 and 30 wt%, respectively (Fig. 2b,c). These trends indicate that adequate ionomer content is necessary to support the increased number of active sites associated with higher catalyst loadings, as the ionomer facilitates proton transport between these sites. However, excessive ionomer content can obstruct electron transfer pathways within the catalyst layer and at the interface between the catalyst layer and PTL, as well as block the catalyst’s active sites, both leading to a reduction in current density [23]. In the catalyst loading range of 0.5–1.5 mg_PtRu cm^–2, the overall cell performance improved with increasing catalyst loading at the optimum ionomer content.

Fig. 2.

(a–c) Polarization curves and (d–f) Nyquist plots of unit cells for electrochemical toluene hydrogenation with cathode catalyst loadings of (a, d) 0.5 mg_PtRu cm^–2, (b, e) 1.0 mg_PtRu cm^–2, and (c, f) 1.5 mg_PtRu cm^–2, as a function of ionomer content. The MCAs were fabricated using carbon paper without an MPL through the CCS method.

EIS was employed to further investigate the influence of ionomer content and catalyst loading on current density. Fig. 2d–f presents the Nyquist plots of MCAs with varying catalyst loadings and ionomer contents. At a catalyst loading of 0.5 mg_PtRu cm^–2 (Fig. 2d), the high-frequency x-intercept, corresponding to R_Ω, remained consistent at ~0.85 Ω across ionomer contents ranging from 10 to 30 wt%. Similarly, R_ct, which is indicated by the difference between the high- and low-frequency x-intercepts, remained nearly constant at ~8.67 Ω. However, at 40 wt% ionomer content, both R_Ω and R_ct significantly increased—by approximately 10- and 1.4-fold, respectively—compared with their values at 10–30 wt% ionomer content. These increases align with the polarization trends shown in Fig. 2a. When the ionomer content was between 10 and 30 wt%, the polarization curves remained similar. However, at 40 wt%, cell performance declined sharply owing to the elevated R_Ω and R_ct. The increase in R_Ω is attributed to excessive ionomer obstructing electron transfer within the catalyst layer and at the interface between the catalyst layer and PTL. The increase in R_ct results from the ionomer blocking the active sites of the catalyst. Both effects contribute to a reduction in the current density. Table 1 lists the current densities measured at 1.95 V and resistances obtained by EIS for various catalyst loadings and ionomer contents. At higher catalyst loadings, the sensitivity of R_Ω and R_ct to changes in ionomer content became more pronounced, leading to greater variability in cell performance. For instance, at a catalyst loading of 1.0 mg_PtRu cm^–2, R_Ω reached a minimum at 20 wt% ionomer content, corresponding to the highest cell current density observed across the cell voltage range investigated (Fig. 2b). Notably, the R_ct did not exhibit a minimum at this ionomer content but instead reached its lowest value at 30 wt%. However, under this condition, R_Ω increased by nearly 70%, from 0.47 to 0.80 Ω, resulting in a reduction in current density. At 40 wt% ionomer content, the R_ct increased sharply because the excess ionomer blocked the active sites of the catalyst. A similar trend was observed at a catalyst loading of 1.5 mg_PtRu cm^–2, where 30 wt% was the optimal ionomer content for the highest current density (Fig. 2c). These results indicate that to achieve the best cell performance, both R_Ω and R_ct must be reduced in a balanced manner by carefully controlling the ionomer content.

Table 1.

Current densities at 1.95 V and cell resistances of unit cells with varying catalyst loadings and ionomer contents

The effect of the catalyst layer formation method and the presence or absence of the MPL on the performance of MCAs in electrochemical toluene hydrogenation was investigated using a catalyst loading of 0.5 mg_PtRu cm^–2. This low catalyst loading was chosen to better emphasize the impact of the catalyst layer formation method and the MPL, rather than the catalyst loading itself. Fig. 3a shows the polarization curves of unit cells with MCAs fabricated using the CCM method, with varying ionomer contents. In this case, the PTL was bare carbon paper without an MPL. Compared with the MCA fabricated using the CCS method (Fig. 2a), the MCA fabricated using the CCM method exhibited higher current densities across the entire voltage range. For instance, the highest current density of 0.427 A cm^–2 was obtained at 1.95 V at an ionomer content of 20 wt%, whereas it was only 0.141 A cm^–2 (at 10 wt%) for the MCA fabricated using the CCS method. The ionomer content that maximized the current density was 20 wt%, which differed slightly from that observed for the CCS method (10 wt%). This difference was attributed to structural differences in the substrate on which the catalyst layer was formed. In the CCM method, the catalyst is primarily deposited on the membrane, ensuring good contact between the membrane and the catalyst layer. By contrast, in the CCS method, the large pore size of the bare carbon paper resulted in significant catalyst penetration into the carbon paper during layer formation, thereby reducing the amount of catalyst in direct contact with the membrane. The enhanced catalyst utilization in the CCM method presumably accounts for the higher current densities compared with those observed using the CCS method. However, when carbon paper with an MPL was employed as the PTL, the overall cell current density was significantly improved, regardless of the catalyst layer formation method, as shown in Fig. 2a and 3c (CCS method) and Fig. 3b and c (CCM method). This improvement is attributed to the MPL providing better electrical and thermal contact between the catalyst layer and PTL by creating a smoother and more continuous interface, as well as facilitating mass transport [24,25]. Notably, because the pore size of the MPL is smaller than that of bare carbon paper [26], catalyst penetration into the carbon paper was significantly reduced when the catalyst layer was formed using the CCS method. This reduction led to catalyst utilization levels comparable to those achieved using the CCM method: the current density measured at 1.95 V was dramatically increased from 0.141 A cm^–2 (at 10 wt%) to 0.441 A cm^–2 (at 40 wt%) when the MPL was introduced (Fig. 2a and 3c). Fig. 4 presents cross-sectional FE-SEM images and EDS elemental mappings of the MCAs fabricated using the CCS method, with and without an MPL. The images clearly show that the PtRu/C catalyst penetrates deeper into the bare carbon paper than into the carbon paper with the MPL. The reason why the optimum ionomer content varies depending on the presence or absence of MPL in MCAs fabricated using the CCS method is not clear. However, it is suspected that the observed difference is related to the low catalyst loading and the impact of the MPL on catalyst layer uniformity. Without the MPL, the catalyst layer formed using the CCS method tends to be thinner and less uniform due to catalyst loss and penetration into the substrate. This reduces the effective contact area between the catalyst layer and the membrane. In such a scenario, adding excessive ionomer could block the limited active sites, leading to a substantial decrease in performance. Furthermore, the non-uniform catalyst layer likely hampers efficient ionomer distribution, which may explain why higher ionomer content has a negative effect. In contrast, when the MPL is present, a thicker and more uniform catalyst layer can be achieved with reduced catalyst loss. This improved layer uniformity allows for better ionomer distribution, making it beneficial to increase the ionomer content, which could enhance overall cell performance. As a result, the performance gap between the MCAs fabricated using the CCS and CCM methods narrowed significantly, as shown in Fig. 3b,c. Despite the incorporation of an MPL, the differences in the polarization curves of MCAs fabricated using the CCS and CCM methods persisted, with the CCM method yielding a higher performance at low current densities. For example, the MCA fabricated using the CCM method delivered a current density of 0.1 A cm^–2 at 1.69 V, whereas the MCA fabricated using the CCS method required 1.72 V to achieve the same current density (Fig. 3d). This indicates that the CCM method has a catalyst utilization advantage over the CCS method, even in the presence of an MPL. Conversely, at high current densities, the MCA fabricated using the CCS method required lower cell voltages than that fabricated using the CCM method. For instance, to achieve a current density of 0.5 A cm^–2, the MCA fabricated using the CCS method required 1.98 V, while the MCA fabricated using the CCM method required 2.0 V.

Fig. 3.

Polarization curves of MCAs fabricated using (a) the CCM method with a bare carbon paper, (b) the CCM method with a carbon paper containing an MPL, and (c) the CCS method with a carbon paper containing an MPL, depending on the ionomer content. (d) Comparison of polarization curves of MCAs fabricated using the CCM and CCS methods with an MPL-containing carbon paper. The cathode catalyst loading for all MCAs was 0.5 mg_PtRu cm^–2.

Fig. 4.

FE-SEM images and EDS elemental mappings of MCAs fabricated using the CCS method: (a) without an MPL and (b) with an MPL.

To better understand the factors influencing cell performance, the HFR was measured using EIS, and iR-corrected polarization curves were obtained. Fig. 5a shows the iR-corrected polarization curves of unit cells with MCAs fabricated via the CCS and CCM methods, using carbon paper with and without an MPL, each at the optimal ionomer content. The configurations are denoted as CCS w/o MPL (10 wt% of ionomer content), CCS w/ MPL (40 wt%), CCM w/o MPL (20 wt%), and CCM w/ MPL (40 wt%). The iR corrected polarization curves for CCS w/o MPL and CCS w/ MPL show that the presence of the MPL improved cell performance: introducing the MPL resulted in a ~90% increase in current density at 1.95 V. In contrast, no significant difference was observed between the iR-corrected polarization curves of CCM w/o MPL and CCM w/ MPL. In the presence of an MPL, at low current densities, the CCM w/ MPL required a lower voltage, while at high current densities, the CCS w/ MPL was more voltage efficient, consistent with the observations in Fig. 3d. Fig. 5b illustrates the variation in the HFR with current density. For the MCAs fabricated using the CCS method, the HFR remained relatively constant across different current densities, and the introduction of an MPL reduced the HFR by ~20%. This reduction is likely due to the improved interfacial contact between the catalyst layer and the membrane, as the MPL provides a larger contact area than that of the bare carbon paper owing to its smaller pore size [25,26]. The MCAs fabricated using the CCM method exhibited HFR values similar to those of CCS w/ MPL at low current densities, but the HFR increased with increasing current density. Because the contact resistance is independent of the current density, this increase is likely related to membrane resistance, potentially caused by membrane dehydration [27,28]. During electrochemical toluene hydrogenation, water crossover occurs from the anode to the cathode, which can induce the HER, a side reaction. If the hydrogen gas generated by the HER is not efficiently removed from the electrode surface, localized membrane dehydration may occur, increasing the membrane resistance. This effect is more pronounced at high current densities owing to increased water crossover [19]. However, the introduction of an MPL mitigated this increase in the HFR, likely owing to the MPL’s hydrophobic and microporous nature, which helps suppress the accumulation of water and hydrogen gas at the interface between the PTL and the catalyst layer [29]. Continuous exposure to toluene may also influence the properties of the polymer ionomer and membrane, potentially contributing to the observed changes in the HFR. However, further studies are required to clarify this effect.

Fig. 5.

(a) iR-corrected polarization curves and (b) HFR values of CCS w/o MPL, CCS w/ MPL, CCM w/o MPL, and CCM w/ MPL as a function of current density. The cathode catalyst loading for all MCAs was 0.5 mg_PtRu cm^–2.

Fig. 6 illustrates the variations in the η_kin and η_mt as a function of current density, determined via the OBA. As shown in Fig. 6a, η_kin increased with rising current density for all configurations. Notably, η_kin was the highest in CCS w/o MPL across the entire range of current densities examined. However, this overpotential was significantly reduced upon the introduction of an MPL, as observed in CCS w/ MPL. This finding reaffirms that the MPL improves catalyst utilization of the MCAs fabricated using the CCS method. Interestingly, MCAs fabricated via the CCM method did not exhibit a marked difference in η_kin, suggesting that the introduction of an MPL to MCAs fabricated by the CCM method does not have a major impact on catalyst utilization. Regarding η_mt, a decrease was observed with the introduction of an MPL, irrespective of the catalyst layer formation method employed (Fig. 6b). MPLs, typically composed of carbon black and hydrophobic carbon materials such as polytetrafluoroethylene, are well known for mitigating water flooding and thus enhancing the mass transport of gaseous reactants to the catalyst layer [24,25,29]. It is suspected that in the PEM electrolyzer used for toluene hydrogenation, the MPL facilitated toluene mass transport by effectively preventing the formation of a water film created by water cross over from the anode to the catalyst layer. This effect of the MPL is expected to become more pronounced at higher current densities, where water crossover is more severe, as reflected in the observed changes in η_mt in MCAs fabricated using the CCM method. However, a slightly different phenomenon was observed in MCAs fabricated using the CCS method. Although the introduction of an MPL effectively reduced η_mt at low current densities (≤ 0.3 A cm^–2), the opposite trend was observed at higher current densities. Typically, mass transport resistance (consequently, η_mt) is expected to increase with current density. Interestingly, at ~0.3 A cm^–2, η_mt of CCS w/o MPL started to decrease. Although the precise mechanisms underlying this observation remain unclear, it is hypothesized that the bare carbon paper with large pores, in conjunction with the catalyst layer, may enhance mass transfer at high current densities, where significant amounts of crossover water and hydrogen gas are generated. Overall, η_mt was lower in the MCAs fabricated using the CCS method than in those fabricated using the CCM method. This difference was attributed to the higher porosity of the catalyst layer formed using the CCS method, compared with that formed using the CCM method.

Fig. 6.

Kinetic and mass transport overpotentials (η_kin and η_mt) of CCS w/o MPL, CCS w/ MPL, CCM w/o MPL, and CCM w/ MPL as a function of current density. The cathode catalyst loading for all MCAs was 0.5 mg_PtRu cm^–2.

Fig. 7 shows the current efficiency for electrochemical toluene hydrogenation as a function of the catalyst layer formation method and the presence or absence of an MPL. At a given current density (0.10 A cm^–2), where mass transport resistance is not severe, MCAs fabricated using the CCS method demonstrated higher current efficiencies compared with those fabricated using the CCM method. The presence or absence of the MPL had no significant effect on current efficiency, suggesting that the porosity of catalyst layer plays a crucial role in determining the efficiency. In a zero-gap cell for toluene hydrogenation, the water crossing from the anode to the cathode reaches the cathode catalyst layer before encountering the MPL. This water permeates the catalyst layer and subsequently contacts the MPL. At low current densities, the facile passage of water through the catalyst layer appears to be more critical than its passage through the interface between the MPL and the catalyst layer. The catalyst layers formed using the CCS method seems to perform more effectively in this regard, offering an advantage over the CCM-fabricated catalyst layer. Although the MPL contributed to an increase in absolute current density, its effect on current efficiency was not significant. The underlying cause of this behavior remains unclear and requires further investigation.

Fig. 7.

Current efficiency for toluene hydrogenation at 0.10 A cm^–2, (a) CCS w/o MPL, (b) CCS w/ MPL, (c) CCM w/o MPL, and (d) CCM w/ MPL. The cathode catalyst loading for all MCAs was 0.5 mg_PtRu cm^–2.

CONCLUSIONS

In this study, we investigated the effect of fabrication conditions on the performance of MCAs in a PEM electrolyzer for electrochemical toluene hydrogenation. Our findings revealed that the performance of MCAs is significantly influenced by catalyst loading, ionomer content, fabrication method, and the introduction of an MPL. The optimal ionomer content is critical for balancing proton transport and active site accessibility, with excessive ionomer content leading to reduced current density owing to the blockage of active sites and electron pathways. The incorporation of an MPL into MCAs fabricated using the CCS method enhances catalyst utilization, thereby reducing the performance gap with MCAs fabricated using the CCM method. This enhancement was achieved by improving interfacial contact and reducing catalyst penetration into the PTL. MPLs also effectively mitigate the mass transport resistance, particularly at higher current densities, by preventing the formation of water film on the electrode surface. Additionally, the η_mt values of CCS-fabricated MCAs were distinctly lower at high current densities, likely owing to better mass transfer through pores in the catalyst layer. The higher porosity of the CCS-fabricated MEAs also led to a higher current efficiency for toluene hydrogenation than the CCM-fabricated MCAs. These findings indicate that optimizing the CCS method and incorporating an MPL can enhance MCA performance, particularly for toluene hydrogenation.

Article information Continued